I run a structured interview before writing a single word, usually 30 to 45 minutes, recorded. I listen for three things: how they structure an argument, what words they repeat, and where they get genuinely animated. That last one is where the real voice lives. From there I build a voice reference doc with sentence rhythm, phrases they'd never use, and go-to analogies, then I start drafting. After working with multiple founders, I can usually nail the voice on the first draft. Clients stop making edits by week three.

Both, and I think the distinction matters less than people make it. Strategy without execution is a deck no one reads. Execution without strategy is content that doesn't compound. My edge is that I can run the brief, the interview, the draft, and the performance review in the same week without handing off and waiting. That said, if I had to pick one strength it's strategy. A bad strategy can't be edited into a good one.

I start with what the business needs to move: pipeline, retention, demo requests, whatever. Then I work backwards. What content directly influences that, what builds the audience that eventually influences that, and what content is just habit. Anything in the third bucket gets cut. If a format or channel hasn't shown traction in 60 to 90 days, I don't optimize it. I replace it with a hypothesis worth testing.

Research and gap analysis: Claude. Interview transcripts: Granola to extract key quotes and themes. First draft: a custom prompt built around the client's voice document, targeting around 70 to 80% quality. My job from there is the last 20%, adding the opinion, the specific detail, the line that makes someone stop scrolling. The whole chain is faster than a traditional workflow and the output is stronger because I'm editing, not starting from a blank page.

Two things. First, prompt quality determines draft quality. Vague prompts produce generic output, so I invest heavily in the brief and the voice doc before touching AI. Second, I run a final edit pass specifically looking for AI tells: over-explained transitions, fake enthusiasm, anything that hedges instead of commits. If I wouldn't say it in a real conversation, it doesn't survive the edit. The goal is for no one, including the client, to tell where the AI stopped and I started.

Every content pillar gets mapped to a business outcome before anything publishes. LinkedIn thought leadership ties to inbound DMs and connection requests from the right buyers. Newsletter ties to trial signups or demo requests tracked via UTM. Blog ties to organic traffic to product pages. I track a short KPI list per client, usually three or four metrics, and review monthly. If content is moving impressions but not the business metric, that's a strategy problem, not a distribution problem, and I treat it that way.

I care about qualified inbound, email open rate among new subscribers, and content-influenced pipeline. I mostly ignore raw follower count, total impressions, and likes. The test I apply to every metric is simple: if this number doubles, does the business notice? If the answer is no, I stop tracking it.

I say so once, clearly, with reasoning. Here is what I would recommend and here is why. If you want to go a different direction, I will execute it well, but I want you to know my concern first. Then I let them decide and deliver either way. What I don't do is silently comply and let a bad strategy ship without flagging it. That isn't being client-friendly, it's being conflict-averse, and it costs clients the results they hired us to deliver.

B2B SaaS, GTM tools, marketing technology, finance, fitness, and fintech. In those I can move from brief to draft without a research sprint because I already understand the buyer, the terminology, and the competitive dynamics. For a new industry I'm transparent about the ramp, usually two to three weeks to get the vocabulary and audience perspective right. What doesn't change across industries is the strategy: who is the reader, what do they fear, and what decision is the content trying to move.

Templates and systems, not heroics. I do not rely on having a good day to produce good work. Brief formats, voice documents, prompt libraries, and QA checklists make quality the default output rather than the exception. Speed comes from removing decisions, not cutting corners. When I know exactly what a piece needs to do, who it is for, and what voice it is in before I start, the draft is fast and the edit is short.